power tube for drivers

[mashaffer]

Would there be any advantage in using what is normally an output tube as a driver? Just wondering if the ability to run them very conservatively would improve distortion over a more traditional driver tube.

mike

[SY]

The principal advantage (in selected cases) is greater max plate voltage and hence larger swing. The more common reason is an illogical gut feeling that a power tube ought to make the sound more, well, powerful.

[mashaffer]

Thanks SY.

mike

[lndm]

Using a power device as a driver has both good and bad points. Driving the output stage into grid current (as happens from time to time) likes a low output impedance from the driver stage (and high current). However, most power devices are low gain, especially old ones. It is when driving the output stage deep into cut off (grid negative) that the low mu driver may itself go up into grid current, upsetting the third last stage.

[Dave Cigna]

I don't think there's a clear line. What one person might call a tiny output tube another might call a beefy driver. The fat half of a 6EM7 makes a very nice 1W output or a very nice driver for a 300B. I suppose it depends on what you're trying to drive.

A high mu transmitting tube like a 211 pretty much requires a small power tube to drive it. The Miller capacitance of any triode requires a little power, especially the older directly heated triodes. A 2A3 can be driven by 1/2 6SL7, but I found that a triode strapped 2A5/42 works better.

The old RCA tube manuals recommended driving PP 42's with a single triode strapped 42 and phase splitting interstage.

The so-called Sakuma method uses 'like drives like;' a 211 to drive a 211, for example. The driver runs at lower signal levels so contributes less distortion than the output tube. Some people claim it works very well, though I've never tried it.

What are you trying to drive?

-- Dave

[rdf]

I played around for a while with an EL84 driving another EL84. Found it way too microphonic to be of use and didn't pursue it. Something like a 4-5 second delay with the Russian clones. A JJ or EH were 'better' if 2-3 seconds counts. I think Pete Millett found the same issue with the EL34 in one of his preamp designs.

[mashaffer]

Thinking in generalizations right now but I kind of had in mind the keyboard amp idea I posted earlier (Transformer coupled, fixed bias, PP 6CA7 connected UL). My initial idea was 6SN7 as the driver but thought that a small power triode (or strapped pentode) might have some advantages. For obvious reasons we want to watch out for microphonics but I want to be sure that the output reaches overload long before the previous stages.

mike

[lndm]

A smaller triode will still drive your o/p well, no fear. Just remember the o/p stage's input capacitance (as Dave Cigna said) for good hf xtn. The 6sn7 is famous as a driver, anyway, and you could always play with xfmr ratios to find out for yourself.

[Dave Cigna]

Now I remember ... 6AS7's (pentode or UL) are pretty easy to drive. A 6SN7 would do. OTOH, a good candidate might be the 6DN7. It's a dissimilar dual triode with an octal base. One side is not too much different from a 6SN7, though with lower ratings. The other side is a small power triode with max plate dissipation of 10W. I think both sides sound very good. It even has the same pinout as the 6SN7, so you might be able to experiment with both without too much rework.

-- Dave

[EC8010]

Of course you can use a power valve to drive another power valve. It may allow you to drive the output valve (or valves) into Class A2 and increase the efficiency of the output stage. However... ...try doing the sums. You'll be amazed at how quickly required driver power supply current rises and squanders the efficiency gain in the output stage.

What is really being discussed here is the ratio between a design's anode dissipation of its output valves and its drivers. Traditional designs used very high ratios (ECC83 driving EL34), whereas contemporary designs use lower ratios in order to deliver clean current into Miller capacitance and output stage grid current (for whatever reason).